Posted on 04/08/2011
A couple of weeks ago, you might have seen a lot of headlines regarding the spike in the volume of the First Trust ISE Global Copper Index Fund (CU). When a fund such as CU rallies 3.9% in one day, and does so on a day when its own volume of 4.5 million shares is only about 100 times its average daily volume, it is human nature to presume one might have something to do with the other. It is important, we feel, to examine both the perspective of the "chicken" (price) and the "egg" (volume) in cases such as this however. Could volume spikes cause tracking errors in ETFs? Surely that is possible, it has happened before (Natural Gas Funds trading well away from "NAV" while new creations were halted a few years ago come to mind), but such events have been few and far between, and generally the result of an inability for the creation/redemption process to operate as designed. We'd like not to assume that material net inflow to a fund would cause prices to change in a manner which they wouldn't otherwise generally have changed.
On March 23rd, the day that CU experienced a high trading volume, it ended up being a pretty good day for Copper. Not just a good day for this particular Copper Equity Fund, but a good day for just about anything copper-related. When the equity market opened at 9:30am EDT Copper Futures had already moved materially higher. May Copper Futures (HG/K1) would rally 2.8% on the day, something that in itself would suggest a good day for the "chicken," regardless of what transpired on the "egg" front. As mentioned in an IndexUniverse.com article that is aptly titled, ETF Fund Flows: CU's Assets Double, the First Trust ISE Global Copper Fund's assets did double in one session last week. However, so long as liquidity is present within the securities that CU must represent, volume in the ETF shouldn't be problematic. The real measure to look at is the volume/liquidity of the underlying securities. The market makers can create as many shares of the ETF as they need, as long as they can deliver the underlying basket of securities. There are many ETF's that are very thinly traded, but since the underlying securities are very liquid you can do 10x the average volume. CU's top holdings are all names with ample liquidity.
Many of CU's top holdings are foreign listed, for simplicity's sake we looked a little closer at the trading history of the two largest US listed holdings of CU: Rio Tinto [RIO] and Freeport McMoran [FCX]. As you can see from the graphic below, while trading volume over the past 3 days (since CU was recommended) was a bit above the average volume over the past 12 months in both RIO and FCX, it was only meagerly so, and in both cases these stocks had higher volume days already in the month of March. In other words, while shares of CU may have seen AUM rise sharply over the course of just a few days, this was not the case for the stocks held within the fund.
Certainly shares of CU were a source of new demand for securities such as RIO, FCX and others last week; demand that cumulatively outmeasured net new supply by evidence that prices rose when all was said and done. However, even if every single retail order in CU last Wednesday was a "buy at the market" order, and all 4.5 million shares of volume equated to $144 million of net new demand for this global Copper portfolio, it still makes a rather meager dent in the daily flow for something like RIO or FCX. For instance, 5% of $144 million is $7.2 million, and that would roughly be the amount of new demand moving toward FCX or RIO as a result of CU buy orders. FCX has a market cap of nearly $52 Billion and handles $7.2 million worth of trading about 23 times over in the course of a "normal" day of trading. Said another way, FCX shares change hands at a rate of about $7.2 million every 17 minutes.
If we can accept that massive volume for a fund like CU does not necessarily equate to "massive" volume for its underlying holdings, than the next logical test is to make sure that the fund is still trading at a price that represents the "sum of its parts". We would explore that by illustration of the funds "Bid/Ask Mid-Point" at the end of the day, compared to the fund's Net Asset Value (NAV) per share at the end of day. The difference between the two, divided by the NAV (per share), would tell us if a fund is trading at inflated prices (premium) relative to its NAV, or deflated prices (discount). This month we've seen shares of CU stay well within +/- 1% of NAV, at least at the close of day, and closed Friday at a 0.23% premium. For perspective, shares of the iShares MSCI Emerging Markets Fund [EEM], which is a nearly $40 Billion fund, trades at a 0.15% premium currently, and during the 4th quarter of 2010 varied from as much as 0.94% premium to a 1.32% discount to NAV. (Source: www.ishares.com) Even the iShares MSCI EAFE Index [EFA] traded to a 1.25% premium at one point last November.